Shulou(Shulou.com)06/01 Report--

Preface

The high availability architecture is basically standard for Internet services, and both application services and database services need to be highly available. For a system, it may contain many modules, such as front-end application, cache, database, search, message queue and so on. Each module needs to be highly available in order to ensure the high availability of the whole system. For database services, high availability may be more complex, and services available to users are not only accessible, but also need to be guaranteed to be correct, so the high availability scheme of the database has always been a hot topic. Today, on a variety of high availability programs, talk about some personal views, if there are any mistakes, please correct them!

The first master-slave synchronization architecture

This kind of architecture is commonly used in start-ups, and it is also convenient for further expansion.

This architecture features:

1. Low cost, quick and convenient deployment

2. Separation of reading and writing

3. It can also reduce the pressure of reading the library by increasing the number of users in time.

4. Single point of failure of the main library

5. Data consistency problem (caused by synchronization delay)

The second Mysql+DRBD architecture

Through the DRBD replication mode based on block block, the double master failover can be carried out quickly, and the problem of single point of failure of the main database can be solved to a great extent.

This architecture features:

1. High availability software can use Heartbeat, which is fully responsible for the management of VIP, data and DRBD services.

2. After the master failure, the slave database can be switched automatically and quickly, and the slave library can still synchronize with the new master database through VIP.

3. Separation of read and write is also supported from the library, which can be implemented by middleware or program.

The third Mysql+MHA architecture

At present, MHA should also be a relatively mature and common scheme in the high availability scheme of Mysql. It was developed by the Japanese. In the process of mysql failover, MHA can achieve fast and automatic switching operation and maximize data consistency.

This architecture features:

1. Installation and deployment is simple and does not affect the existing architecture.

2. Automatic monitoring and failover

3. Ensure data consistency

4. Failover mode can be manual or automatic multi-directional selection

5. Wide range of adaptation (applicable to any storage engine)

The fourth Mysql+MMM architecture

MMM, or Master-Master Replication Manager for MySQL (mysql Master replication Manager), is a scalable suite of scripts for monitoring, failover, and management of mysql master replication configurations (only one node can be written at any time). This suite can also read load balancing from any number of slave servers based on the standard master-slave configuration, so you can use it to launch virtual ip on a group of servers that reside in replication In addition, it has scripts that implement data backup and resynchronization between nodes.

MySQL itself does not provide a replication failover solution, through the MMM scheme to achieve server failover, so as to achieve the high availability of mysql.

The features of this scheme are:

1. High security, high stability and good scalability

2. At least three or more servers are required

3. For double masters (higher requirements for master-slave reproduction)

4. The separation of reading and writing can also be realized.

The fifth official mysql cluster scheme

The cluster high availability scheme officially launched by Mysql is difficult to evaluate because I have not used it.

An explanation taken from the Internet

MySQL Cluster consists of a group of computers, each running a variety of processes, including MySQL servers, NDB Cluster data nodes, management servers, and (possibly) specialized data access programs.

Due to the complexity of the MySQL Cluster architecture, deployment is time-consuming (it usually takes several hours for DBA to complete the build), while relying on MySQL Cluster Manager can be done with only one command, but MySQL Cluster Manager is paid for. And industry veterans believe that NDB is not suitable for most business scenarios and has security problems. Therefore, the number of users is relatively small.

If you are interested, you can study the official document: https://dev.mysql.com/doc/refman/5.7/en/mysql-cluster-install-linux.html

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.